Artist claims Meta and OpenAI violate copyright by training AIs

Comedian and actress Sarah Silverman has sued Meta and OpenAI for copyright infringement. The artist is the company’s artificial intelligence platforms Large Language Model Meta AI (LLaMA) and ChatGPT, illegally trained using excerpts from one of his books.

The lawsuit was filed in San Francisco, California, and also represents the claims of writers Christopher Golden and Richard Kadrey. The duo also accused the AI ​​systems of providing duplicate summaries of their books in their entirety.

Evidence collected by lawyers shows quotations that appear to have been directly copied from publications, bedwetter (from Silverman), sand man thin (Kadrey) and ararate (Gold). When requesting summaries of books in ChatGPT, the app only provides all chapters written by authors. The books also appear in searches by Meta’s conversational AI system, LLaMA.

The legal representatives of the authors claim that the sources the platforms refer to are illegal. The lawsuit argues that the data was obtained from Library Genesis, a search engine for books and scholarly articles, Bibliotik (which makes books downloadable via torrent), Z-Library, and ThePile (two shadow libraries).

A shadow library, or “shadow library” in literal translation, is a publication repository that offers free copyright content that is usually no longer available in stores. However, these collections also contain copyrighted works that are illegal to download.

In court, Sarah Silverman, Christopher Golden and Richard Kadrey they say they “do not allow their copyrighted books to be used as educational material”. The lawsuit also accuses Meta and OpenAI of negligence, illegal enrichment and unfair competition, among others.

Other side

sites Business Content This Border He contacted OpenAI and Meta, but the companies did not comment on the matter.